Datasets:

---

Hafftka

/

michael-hafftka-catalog-raisonne

like 31

Hello,

I uploaded a dataset of ~3,780 images (≈40 GB) to the Hugging Face Hub:

https://huggingface.co/datasets/Hafftka/michael-hafftka-catalog-raisonne

The dataset structure is:

images/
metadata.csv
metadata.parquet
README.md

metadata.csv and metadata.parquet both contain the following columns:

catalog_number,file_name,artist,title,year,medium,dimensions,collection,copyright_holder,license,view

The file_name column contains relative paths such as:

images/<filename>.jpg

I initially had a dataset_info block in the README, but after encountering viewer errors I removed it and also deleted dataset_infos.json.

However, the Dataset Viewer page still shows:

ConfigNamesError
AttributeError: 'str' object has no attribute 'items'

Traceback indicates the error occurs when parsing dataset_info["features"], even though that block is no longer present in the README.

Could the viewer metadata cache be stale, or does the indexing pipeline need to be reset for this repository?

Thanks for any help.